You uploaded your audiobook to ACX, waited for the review, and got the rejection email: RMS loudness outside the acceptable range. Maybe your files are too quiet, sitting at -30 dBFS when ACX needs -23 to -18 dBFS. Maybe they're too loud, pushed past -18 dBFS in a well-intentioned attempt to sound professional. Either way, the fix is the same: understand what ACX measures, why your files failed, and how to land in the sweet spot.[1]
This guide focuses specifically on loudness and RMS rejections, the single most common reason audiobooks fail ACX's automated quality check. For the full technical context on loudness measurement, see the complete loudness guide. If you're dealing with other rejection types, see the complete ACX rejection troubleshooting guide.
What Does RMS Loudness Actually Mean for Audiobooks?
RMS stands for Root Mean Square, a mathematical method of calculating the average power of an audio signal over time. Unlike peak measurement, which captures the single loudest moment, RMS represents the overall perceived loudness of your entire chapter file.[2]
ACX requires every chapter file to have an RMS loudness between -23 and -18 dBFS. This five-decibel window ensures that listeners experience consistent volume across chapters within your audiobook and across different audiobooks on the Audible platform.[1]
How Does ACX Measure RMS?
ACX's automated check measures the full-file RMS of your submitted MP3. This means it calculates the average loudness across the entire duration of the chapter, including speech and silence. This has important implications:[2]
- Long pauses lower your RMS. A chapter with frequent pauses or long silences will measure quieter than a densely narrated chapter of the same duration.
- Dynamic narration can shift your average. A narrator who whispers and shouts within the same chapter will have a different RMS than one with consistent delivery.
- The measurement is on the final MP3, not your project file. MP3 encoding can shift levels slightly, so always verify the delivered file.
The measurement uses a standard RMS calculation with no frequency weighting. This differs from LUFS (Loudness Units Full Scale), which applies a K-weighting curve that matches human perception per ITU-R BS.1770. ACX uses RMS, not LUFS, though the two measurements are related for spoken word content and typically track within 1–2 dB of each other.[3]
Why Was Your Audiobook Rejected for Loudness?
There are two directions a loudness rejection can go, and the causes are different for each.
What Causes RMS Too Low (Below -23 dBFS)?
This is by far the more common rejection. Most home narrators record conservatively to avoid clipping, producing raw recordings that land around -30 to -40 dBFS. Without proper gain staging and mastering, these files will always fail.
Common causes:
- Recording too quietly. Microphone too far from mouth, input gain too low, or both.
- No gain adjustment applied. The raw recording was exported without loudness processing.
- Excessive dynamic range. Whispered passages pull the average down even when louder sections are in range.
- Too much silence. Long pauses between paragraphs or sections dilute the average loudness.
- Noise gate too aggressive. A gate that silences too many quiet speech passages removes energy that contributes to RMS.
What Causes RMS Too High (Above -18 dBFS)?
Less common but still a real problem, especially among narrators who learned mastering from music production guides. Music targets are typically -14 to -8 LUFS, significantly louder than ACX's range.
Common causes:
- Over-compression. Heavy compression raises the average level by squashing dynamics.
- Excessive normalization. Normalizing to -16 or -14 dBFS pushes audiobooks too loud.
- Incorrect target. Using LUFS targets from podcast or music specs instead of ACX's RMS range.
- Processing chain designed for music. A mastering chain optimized for music loudness doesn't transfer to audiobooks.
How Do You Fix RMS Too Low?
This process assumes your recording is clean and edited but too quiet. If you also have noise floor issues, address those first because gain adjustments amplify background noise.
Step 1: Measure Your Current RMS
Open your chapter file in your DAW and measure the full-file RMS. In Audacity, select all audio (Ctrl+A) and use Analyze → Contrast or the ACX Check plugin. Note the current value. You need to know how far you are from the target.
Step 2: Apply Gentle Compression
Compression reduces the gap between your loudest and quietest moments, making gain adjustment more effective. For spoken word, use conservative settings:[4]
- Ratio: 2:1 to 3:1 (never higher for narration)
- Threshold: Set so the compressor engages on louder passages, typically -20 to -18 dBFS
- Attack: 15–25 ms (fast enough to catch transients, slow enough to preserve natural speech onset)
- Release: 200–500 ms (matches natural speech rhythm)
- Makeup gain: Don't apply yet. Handle gain in the next step
Compression alone won't fix the problem, but it narrows the dynamic range so that the subsequent gain adjustment brings the whole signal into range rather than just the peaks.
Step 3: Adjust Gain to Target -20 dBFS
Target the middle of ACX's range: -20 dBFS RMS. This gives you 2 dB of margin in each direction, accounting for measurement variation between tools.
In Audacity: Effect → Loudness Normalization, set to -20 dBFS with RMS mode selected. (Audacity 3.x supports both LUFS and RMS modes, so make sure RMS is selected for ACX compliance.)[5]
After gain adjustment, re-measure your RMS. If you're between -21 and -19 dBFS, you're in an excellent position.
Step 4: Limit True Peaks
Raising gain pushes peaks higher. After gain adjustment, your true peaks may exceed -3 dBFS. Apply a true-peak-aware brickwall limiter set to -3.1 dBFS or lower.[6]
Use a true-peak limiter, not a standard sample-peak limiter. Standard limiters miss inter-sample peaks that ACX's check will catch. FabFilter Pro-L, iZotope Maximizer, and Waves True Peak Limiter all handle this correctly. If using Audacity, its built-in limiter is sample-peak only, so set it to -3.5 dBFS to compensate.
Step 5: Re-measure Everything
After processing, verify your final MP3 against all specs:
- RMS between -23 and -18 dBFS ✓
- True peak below -3 dBFS ✓
- Noise floor below -60 dBFS ✓
Always measure the MP3 output, not your project file. MP3 encoding can shift both RMS and peak levels slightly.
How Do You Fix RMS Too High?
The fix for too-loud files is more straightforward but requires careful attention.
Step 1: Reduce Overall Gain
Attenuate your file to bring RMS down to approximately -20 dBFS. In most DAWs, this is a negative gain adjustment. Measure after adjustment to confirm you're in range.
Step 2: Check Your Compression Settings
If you applied heavy compression during mastering, reduce the ratio. A 2:1 ratio is appropriate for spoken word. Anything above 4:1 is compressing too aggressively and artificially raising your average level.[4]
Step 3: Review Your Processing Chain
If your mastering chain was designed for music or podcasts, start over with audiobook-appropriate settings. Audiobook mastering needs less compression, less limiting, and a lower loudness target than other formats. See the complete audiobook mastering guide for the correct signal chain.
How Do You Prevent Loudness Rejections?
The best fix is avoiding the problem. Here's how to set up your workflow to consistently land in the -23 to -18 dBFS range.
Record at the Right Level
Set your input gain so that normal narration peaks around -12 to -6 dBFS on your recording meter. This gives you headroom for emphasis and emotion while keeping your raw recording loud enough that minimal gain adjustment is needed later.
Use a Consistent Monitoring Setup
If you monitor while recording, keep your monitoring level consistent between sessions. If your headphone volume changes, your vocal delivery unconsciously adjusts with it, creating inconsistencies between chapters.
Process All Chapters Identically
Use the same mastering settings for every chapter in your audiobook. Batch processing ensures this automatically. Chapter-to-chapter loudness inconsistency is a separate rejection reason that often accompanies RMS failures.
Verify Before Submitting
Check every chapter's RMS before uploading. The ACX audio requirements guide covers all eight specs in a systematic order. Ten minutes of verification can save days of rejection and resubmission.
How Do RMS, Peaks, and Noise Floor Interact?
Loudness adjustment doesn't happen in isolation. Changing your RMS affects your peaks and noise floor:
- Raising RMS raises peaks. If your peaks were at -6 dBFS and you add 6 dB of gain, your peaks hit 0 dBFS, causing clipping. Always limit after gain adjustment.
- Raising RMS raises noise floor. A noise floor at -65 dBFS in your raw recording becomes -59 dBFS after 6 dB of gain, above ACX's -60 dBFS threshold.[2]
- Compression changes the relationship. By narrowing dynamic range, compression means less gain is needed to hit the target, which means less noise amplification. This is why the mastering order matters.
The correct processing order is always: noise treatment → compression → gain adjustment → peak limiting → format conversion → verification. This order minimizes the interaction between adjustments. For a detailed walkthrough, see the complete audiobook mastering guide.
When Should You Consider Automated Tools?
If you've tried the steps above and your files still aren't landing in range, or if you're processing a large project with dozens of chapters, the manual approach becomes time-consuming and error-prone. Every chapter needs identical treatment, and one missed setting can mean another rejection.
ChapterPass applies the complete mastering chain (gain adjustment, compression, peak limiting, noise management, format conversion, and silence padding) automatically and consistently across all chapters. The deterministic processing ensures every file meets ACX's specs without manual parameter tuning.
Frequently Asked Questions
What RMS target should I aim for?
Target -20 dBFS. This is the center of ACX's -23 to -18 dBFS range and gives you 2 dB of margin on each side.
Can I use LUFS instead of RMS for measurement?
ACX measures RMS, not LUFS. For spoken word, the two measurements are typically within 1–2 dB of each other, but they're not identical. Use an RMS meter for ACX verification.[3]
Why does my RMS change after MP3 encoding?
MP3 is a lossy compression format that removes audio data. The encoding process can slightly shift both RMS and peak levels. Always measure your final MP3 output, not your project file.
How do I fix inconsistent RMS between chapters?
Process all chapters with identical settings using batch processing. If chapters were recorded at very different levels, normalise each to the same target before applying the rest of your mastering chain.
Stop guessing at loudness levels. ChapterPass measures, adjusts, and verifies RMS for every chapter automatically.